blob: 66f192c414ccf34b0d707f38091f261f53cd258a (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
|
Optional types
==============
Optional types/variables could go at two locations:
* At the type, such as: `?String name`
* At the variable/field: `String name?`
Note that the latter does not support optionals in generic parameters.
The choice also affects changeability/mutability qualifiers.
Syntax examples:
# At type
?String name
?Writer! wr
?String last_url!
?List Int items
# At variable
String name?
Writer! wr?
String last_url!?
List Int items?
# or should `?` come before `!`
String last_url?!
Optional types in type-generic elements
---------------------------------------
It would be tricky (but not impossible) to support optional types in type
parameters of generic types.
The problem is that the `None` value can appear both at the type parameter
level (`T?`, where `T` is of an unknown type), and at the usage level
`List ?String`.
This could possibly be implemented by counting the "level" where the `None`
value originates from. But is that possible to do with multiple levels of
generics? Such as the following:
StringMap List List String
I think it should be possible. I don't think there could be more than a 2-
level ambiguity, because each nesting level of the generic type has its own
totally separate generic storage slot.
So it should be sufficient to have 2 different `None` values:
* `None`
* `GenericNone` / `EmptySlot` / `BlankSlot`
- This one could have be implemented separate value,
in addition to the `NULL` value. Perhaps `1` (or all ones).
This would also mean that the varstate tracking needs to be aware of 3 cases:
EmptySlot, None, and value-present.
Note that an EmptySlot means that the generic value does not exist at all,
so it isn't None it not even valid (similar to an unassigned variable).
|
